Geiling, J M et al 2016 A Spatial Distribution Study of Faunal Remains from Two Lower Occupation Levels in El Mirón , Cantabria, . Papers from the Institute of Archaeology, 26(1): Art. 4, pp. 1–16, DOI: http://dx.doi.org/10.5334/pia-477

RESEARCH PAPER A Spatial Distribution Study of Faunal Remains from Two Lower Magdalenian Occupation Levels in El Mirón Cave, Cantabria, Spain Jeanne Marie Geiling*, Lawrence G. Straus†, Manuel R. González-Morales* and Ana Belen Marín-Arroyo*

Human behaviour can be reconstructed by analysing specific activities and­campsite organization using spatial analysis. The dense occupation layers of the Lower ­Cantabrian Magdalenian in the Northern Spain reveal varied aspects of Upper Palaeolithic lifeways, including evidence of specific localized activities. The outer vestibule of El Mirón cave has a particularly rich and intact Lower Magdalenian occupation horizon, Levels 15–17. The excavations in the outer vestibule “Cabin” area of the site revealed excellent bone preservation. Artefacts and faunal remains were individually recorded and sediments water-screened to yield a large sample of archaeological finds and spatial data. Zooarchaeological analysis provided the taxonomic, anatomic and taphonomic determination of the faunal individual finds. Smaller animal remains were categorized and counted; special attention was given to the identification of anthropogenic modifications such as burnt bones or bone flakes. These small refuse items are considered to be useful, in situ indicators of localized activities. The spatial distribution analysis of this dense and complex palimpsest of El Mirón Lower Cantabrian Magdalenian layers required GIS based methods including density analysis, heatmaps and cluster analysis. Based on the spatial distribution of Level 15 and 16 faunal remains, different activity areas were identified comprising , working and dropping zones. These results imply the deliberately segregated use of space within the Lower Cantabrian Magdalenian site area, in which bone-processing activities played a central role.

Introduction to changing environments and climatic con- There is a broad interest in the study of ditions among Palaeolithic hunter-gatherers activities through the reconstruction of (Binford, 1978; Nakazawa et al., 2009). Analysis ­behaviour patterns, variability and adaptation of specific activities can be used to understand

* Instituto International de Investigaciones † Department of Anthropology, University of Prehistóricas de Cantabria, New . Albuquerque, USA Universidad de Cantabria Corresponding author: Jeanne Marie Geiling Santander, Spain ([email protected]) Art. 4, page 2 of 16 Geiling et al: A Spatial Distribution Study of Faunal Remains from Two Lower Magdalenian Occupation Levels in El Mirón Cave, Cantabria, Spain

Palaeolithic human behaviour. It is widely presumably a frequent activity. Although recognized that zooarchaeological remains plant foods can play a role in the diet, under provide the means to reconstruct Palaeolithic glacial conditions the basis of subsistence behaviour (Bartram et al., 1991) and campsite was mainly fat and meat. A broad spectrum structures (O’Connell et al., 1991). Activities of activities based on animal remains and of interest for the exploration of campsite their subsequent spatial distribution can be organisation include butchering, bone mar- used to define activity zones such as roast- row extraction, grease rendering or bones ing, meat-, marrow- and grease-processing or used as fuel. Based on ethnographical studies, discard (Henry, 2012). Distinct accumulating distributions of tiny bone fragments in camps agents ( versus raptors and terres- may reflect the location of activities generat- trial carnivores), anthropogenic modifica- ing bone refuse (processing, consumption, and tions and preservation conditions can all be discard). The intensity of secondary disposal identified with the help of zooarchaeological activities and taphonomic factors (trampling, investigations (Lyman, 1994; Reitz and Wing, post occupational carnivore/scavenger activi- 2008; Straus, 1982; Villa et al., 2004). ties) can also give insights into subsequent Stratigraphically distinguishing between actions (Bartram et al., 1991, 137). single occupational events and activities is Spatial distribution analysis is based on the difficult or impossible. It is generally assumed three-dimensional relationships of archaeo- that anthropogenic layers, such as archaeo- logical remains (Schiffer, 1996). Scrutiny of logical levels within thick cave deposits, con- object distributions within human occupa- sist of recurring occupations and activities, tion layers is a useful for reconstruct- commonly known as palimpsests, that are ing intra-site spatial organisation and use conditioned by the physical features of each (Clarke, 1977). By using these methods cave such as walls, entrance and , archaeologists are able to reconstruct activity blocks, ceiling height, sunlit vs. dark areas areas (Ferring, 1984, 117). These are gener- (Bailey, 2007; Henry, 2012; Schiffer, 1996; ally exemplified by hearth-centered models, Straus, 2008, 1979). The term palimpsest where different types of activities are spa- refers to the superimposition of successive tially distinguishable (Alperson-Afil et al., activities or traces of multiple, overlapping 2009; Audouze and Enloe, 1997; Fladerer activities revealing time-averaged assem- et al., 2002; Goren-Inbar and Alperson-Afil, blages (Henry, 2012). In the absence of evi- 2010; Leroi-Gourhan and Brézillon, 1966; dent structures which segregate space in Nakazawa et al. 2009). According to Binford a tangible way, zooarchaeological studies (1978) three main areas of bone distribution may provide insights into spatial organisa- around a hearth are identifiable: a drop zone, tion and site formation processes such as a seating area and a toss zone. More recently “invisible ” features (Goren-Inbar these basic categories have been widened and Alperson-Afil, 2010; O’Connell et al., to incorporate , production, cura- 1991). Particularly, good bone preservation tion, walking or sleeping areas and multiple can indicate little post-depositional distur- hearths, increasing the evidence for func- bance from natural agencies and make it tional patterning of intra-site spatial use in possible to investigate the evidence of living the Palaeolithic (see also Henry, 2012). floors. According to plant-animal subsistence The identification of Palaeolithic living ratios among worldwide modern hunter- floors relies on the notion that small refuse gatherers (Cordain et al., 2000), diets include items are prone to remain in place to a more than 50% animal-based foods (more greater extent than large ones. Small bone in cold and high-latitude settings such as fractions (generally < 2–3 cm) suffer less Oldest Dryas northern Spain) and there- transportation and are therefore suitable fore the exploitation of faunal remains is markers of activity zones or living floors Geiling et al: A Spatial Distribution Study of Faunal Remains from Two Lower Art. 4, page 3 of 16 Magdalenian Occupation Levels in El Mirón Cave, Cantabria, Spain

(Malinsky-Buller et al., 2011). Additionally thick, dense artefact, manuport and bone they can represent primary refuse as they accumulations make it difficult to differen- are less likely to be displaced during surface tiate individual living floors during excava- cleaning (Schiffer, 1996). Therefore small tion, and therefore they are usually treated as refuse items such as manufacturing debris units and referred to as a palimpsest. General of antler, bone or stone found in association changes in duration and/or frequency of with finished can be helpful in identify- site occupations following the Last Glacial ing activity zones. Maximum (LGM) are indicated by these dis- This paper asserts that reconstructions tinctive, thick LCM archaeological deposits. of activity zones within campsites for dia- It is assumed that the history of occupation chronic behaviour studies are more precise of these sites reveals varied aspects of Upper and complete when they incorporate small Palaeolithic lifeways, while their dense, bone fractions as indicators of primary refuse detailed record provides considerable, albeit areas. Density analyses are used to create difficult to untangle and decipher, evidence plots of spatial distributions to investigate of specific activities. dense archaeological layers and enable the This paper seeks to propose intrasite spa- interpretation of activity zones. tial analyses for the purpose of identifying The Lower Cantabrian Magdalenian (LCM) LCM hunter-gatherer activities in El Mirón occupations found throughout the north- cave based on faunal remains. Combined ern Atlantic region of Spain provide exam- analytical studies of zooarchaeological meth- ples of such dense archaeological layers. ods and spatial distribution analysis using This cultural period, spanning 16-14.5kya, is Geographical Information Systems (GIS) ena- characterised by a specific suite of material bled the first author to achieve these goals. culture remains. Stone tools include bladelet Zooarchaeological and taphonomic studies cores and nucleiform endscrapers as as a inferred the existence of very good preserva- variety of backed elements (especially balde- tion conditions. The identification of activity lets, but also geometric ). Due to zones such as hearth, bone processing areas the favourable preservation conditions in and dumping zones were condensed into Northern Spain, organic remains are com- systematic plots. The spatial relationship of monly found, including antler sagaies with animal remains inside the rich LCM layer varied (but especially quadrangular) cross- reveals some aspects of Palaeolithic behav- sections and bases (mainly single-bevel), iour, reconstructed using the distributions of needles, perforated shells, and regionally zooarchaeological data. and temporally diagnostic red deer scapulae with red deer hind engraved images whose Material and Methods outlines are filled with striations (González- Lower Cantabrian Magdalenian Deposits Morales et al., 2006; Straus and González- In El Mirón Cave Morales, 2012a). Many LCM levels have one El Mirón is a west-facing cave located in the overarching aspect in common, which is the upper Rio Asón valley on the northern edge extraordinary density of occupation layers, of the Cantabrian Cordillera, some 20 km very rich in organic materials, hearth rem- upstream from the modern Atlantic shore- nants, artefacts, etc., as found at Altamira, line, surrounded by 1000-meter high moun- El Castillo, El Juyo, Rascaño, Santimamiñe tains and dominating the confluence of two and El Mirón cave (Outer Vestibule levels 15, tributaries with the Asón (Fig. 1). The cave 16 and 17, Mid-Vestibule Level 312, Inner has been excavated since 1996 under the Vestibule levels 109–116, and Vestibule Rear direction of L.G. Straus and M.R. González- levels 504–505) (González Echegaray and Morales. The site is well dated with 84 radio- Freeman, 1992; Gonzalez-Morales and Straus, carbon dates ranging from the late Middle 2014; Straus, 1992). These extraordinarily Palaeolithic to the early (Straus Art. 4, page 4 of 16 Geiling et al: A Spatial Distribution Study of Faunal Remains from Two Lower Magdalenian Occupation Levels in El Mirón Cave, Cantabria, Spain

Figure 1: The site of El Mirón is located at the northwest slope of Monte Pando (Ramales de la Victoria) in Cantabria, Northern Spain; at the confluence of the rivers Caldera, Gandára and Asón surrounded by 1000m high mountains in the first main chain of the Cantabrian Cordillera. and Gonzalez-Morales, 2012b, 2010, 2007a; level 17: 15.470 ± 240 (GX-24466) and 15.700 ± Straus and González-Morales, 2003). 190 (GX-25853) radiocarbon years before pre- Two main excavation areas are located in sent (Straus and González-Morales, 2012a). the outer and rear sectors of the cave vesti- The zooarchaeological material from these bule (Cabin and Corral areas respectively), layers shows the continuous hunting of both which are connected by a 9 × 1-meter Mid- Cervus elaphus and Capra pyrenaica. Vestibule Trench (Fig. 2). Between 2001 and 2013 additional Lower Magdalenian layers Field Methods were excavated in the smaller “Burial Area” During excavation at El Mirón cave, stand- behind a large engraved block, where a pri- ard methods of Palaeolithic excavations mary human burial covered by sediments were applied. These methods include: three- with specular hematite crystals and abundant dimensional individual recording of lithics red ochre was uncovered (Geiling and Marín- larger than 1 cm, bones longer than 5 cm, as Arroyo, 2015; Straus et al., 2011; Straus and well as of smaller, but field-identifiable spec- Gonzalez-Morales, 2015). For this case study, imens (e.g., teeth, phalanges) with a total only Levels 15 and 16 were chosen, which station theodolite. The excavations were con- correspond to the uppermost part of the ducted in subsquares (50 × 50 cm) by spits 80 cm-thick LCM deposit in the 9–10 m2 Cabin (app. 1–3 cm thick), the sediments were then area of El Mirón cave (Straus and González- water-screened through fine mesh to collect Morales, 2007b). The Magdalenian sequence the smaller finds or “non-identifiable” bones in the Cabin area is dated on the overlying in so called “General Bags” (GB). Spits are Middle Magdalenian level 14: 14.600 ± 190 arbitrary subdivisions of natural levels. Nine (GX-32383), the LCM level 15: 15.010 ± depth measurements were taken for top and 260 (GX-23392); 15.220 ± 300 (GX-23393) bottom heights of each spit’s meter square and level 16: 15.180 ± 100 (GX-23415) centre, edge center and corner. Based on and the upper part of the underlying LCM these, upper (Level 15), middle (Level 16.1) Geiling et al: A Spatial Distribution Study of Faunal Remains from Two Lower Art. 4, page 5 of 16 Magdalenian Occupation Levels in El Mirón Cave, Cantabria, Spain

Figure 2: The two main excavation areas, Cabin and Corral, are connected with a trench and, behind the engraved block, a human burial was excavated. The material from this study originates from the Cabin area. Lower Cantabrian Magdalenian deposits were found in all other excavation areas (plan modified, original outlines made by GIM GEOMATICS company). and lower samples (Level 16.2) were articu- in each direction, for X and Y coordinates lated (L.M. Fontes pers. comm.). respectively. Gathered identified GB data creates more reference points for distribu- Laboratory Methods tion analysis and also increases the density This study will focus on individually plotted of point clouds. The higher concentrated sin- macromammal finds and remains from the gle data points inside a defined unit are, the GBs. Faunal elements were classified taxo- less informative simple scatter plots become nomically and anatomically. GB items were (Oron and Goren-Inbar, 2014). included to increase the usable taxonomic, Heatmaps were created to represent arte- anatomic and taphonomic information. All fact densities. Heatmaps show areas of high identified faunal remains were used to cal- and low point density visually distinguished culate the Minimum Number of Individuals by colour coding. For individual data sets (MNI). GB find categories were bone or teeth counts the procedure points features (rep- fragments smaller than 5 cm, burnt bones, resenting artefacts) inside a specified area; bone flakes, young spongy bones or digested this study used analytical radiuses of 0,2m bones. The final numbers of teeth and bones around each point. The outcome of a kernel specimen vary widely between individual density analysis is a surface of raster cells, and GB data per square (see Table in Fig. 3). illustrating the sum of overlapping radiuses The taxonomically and anatomically around each point. Here, a small cell size of identified GB items were assigned to sub- 0,01m was chosen to create smooth density squares and do not have coordinates. In maps. The highest values within the plots order to make them available for distribu- correspond to 1 and the lowest values cor- tion analysis a program tool was used to plot respond to 0, represented by colour codes, identified GB finds; GB items received coor- red to blue respectively (Baxter et al., 1997; dinates based on an excel-function (Gilead, Silverman, 1986). 2002). The following function was used: Areas with different densities can be RANDBETWEEN(0;50)/100 to give random summarized in clusters (Baxter et al., 1997; numbers according to the two-dimensional Silverman, 1986; Whallon, 1984). To identify subsquare extensions between 0 and 0,50m clusters during analysis two threshold values Art. 4, page 6 of 16 Geiling et al: A Spatial Distribution Study of Faunal Remains from Two Lower Magdalenian Occupation Levels in El Mirón Cave, Cantabria, Spain

                                                                                                                                           Figure 3: Table. Overview of the counts from the individual finds (>5cm) and GB data (<5 cm) by squares for Level 15 and 16. BB = burnt bones, F = bone flakes and YOU = porous bones from very young animals (red represents the highest numbers, blue the lowest). were defined: one 50 % boundary to delimi- Density and cluster analysis are methods tate low from high densities, plus an addi- used to simplify high concentrated data tional 95% boundary to identify hot spots points (Blankholm, 1991; Hodder and Orton, with high concentrations. These clusters are 1976), but also to find patterns and allow considered to be a sample from the gathered comparisons between levels. The term den- raw data and provide useful insights in distri- sity used here refers to “relative density” by bution patterns (Blankholm, 1991; Whallon, comparing quantities of objects per unit 1984). area or surface measured in square meters. The smaller faunal remains (less than 2–5 cm) Consequently density is a relative term, while found within GB are generally not identified it refers to the overall data found within a in terms of taxonomic classification and were given unit. instead counted per subsquare. GBs also contained teeth enamel fragments or small Results bone fractions. Other categories follow spe- The taxonomical identification of animal cific types of anthropological modifications remains from level 15 and 16 revealed that like burnt bones for hearth activities or bone most of the identified taxa (NISP = 1.854) flakes originating from percussion through could be attributed to Cervus elaphus (MNI = intentional bone cavity opening related to 2; NISP = 680; 36,7%) and Capra pyrenaica actions for bone marrow and grease extrac- (MNI = 10; NISP = 1.100; 59,3%). The suc- tion (Costamagno et al., 2013; Manne et al., ceeding spatial distribution analysis of indi- 2006; Outram, 1998). The small fraction cat- vidual items will concentrate on these two egories are of particular interest in identify- species, while taphonomic studies revealed ing Palaeolithic activity zones, and heatmaps an anthropogenic character of accumula- were created for each sublevel by subsquare. tion (Geiling, 2014). The majority of the GB The colour code used for the subsquares content was identified as small bone and ranges also from blue to red, from low to teeth fragments (n = 50.658). Burnt bones high density respectively. (n = 2.051) were also represented. Additional Geiling et al: A Spatial Distribution Study of Faunal Remains from Two Lower Art. 4, page 7 of 16 Magdalenian Occupation Levels in El Mirón Cave, Cantabria, Spain categories are bone flakes (n = 653) indicat- dumping faunal material; in other words, the ing intentional bone processing activities bones of both species were processed and (possibly marrow and grease extraction), discarded in the same areas across the site. and porous bones from young individuals McKellar’s Principle, which states that (n = 304), which indicate the favourable small refuse items are commonly discarded preservation conditions of these archaeologi- in the same area in which they were pro- cal layers. duced, rather than disposed of somewhere The general spatial distribution of all indi- else, was used to analyse the GB data. By vidual faunal remains shows dense archaeo- this principle larger items are more prone to logical accumulations in two areas: one transport, while small in situ refuse, tends to across squares H2, H3 and H4 and another penetrate into the sediment through tram- one in square J3 (Fig. 4-A). The spatial dis- pling (McKellar, 1983; Schiffer, 1983). This tribution of the two main game species allows archaeologists to reconstruct living (Fig. 4-B), Cervus elaphus and Capra floors. ­pyrenaica, show similar concentration across Burnt bones (<2 cm) are smaller than the Cabin area. The close clusters of 50%- unburned residues (Costamagno et al., 2006; and 95%-density borders indicate a rapid Stiner et al., 1995) and generally remain increase in point intensities in the same in place; burnt bones may help to localize areas of higher general densities. The spatial hearths, especially if structural elements cluster of high-density areas for each spe- missing (Alperson-Afil et al., 2009; Goren- cies overlap by more than 50%. The similar- Inbar and Alperson-Afil, 2010). The spatial ity in find provenance indicates that there distribution of burnt bones from GBs is used is no species-specific area for processing or to identify hearths in the LCM levels. Burnt

A)Density of all animal remains: B) Species cluster: Cervus elaphus 50% 95% 010.5 low high Capra naica 50% 95% HI J HI J 4.0 4.0

4 3.5 4 3.5

3.0 3.0

3 2.5 3 2.5

2.0 2.0

2 1.5 2 1.5

1.0 1.0 1 1 N 0 0.5 1 m N 0 0.5 1 m 0.5 0.5 7.07.5 8.08.5 9.09.5 10.0 7.07.5 8.08.5 9.09.5 10.0 Figure 4: The spatial distribution pattern of faunal remains shows: A) generally high den- sities corresponding to hotspots in meter- squares H3, H4 and J3; B) that the spatial distribution clusters of high densities of the two main game species, Cervus elaphus and Capra pyrenaica, largely overlap and indicate similar processing or discarding activities (Grey points mark individual faunal finds). Art. 4, page 8 of 16 Geiling et al: A Spatial Distribution Study of Faunal Remains from Two Lower Magdalenian Occupation Levels in El Mirón Cave, Cantabria, Spain bones found at El Mirón LCM layers corre- The spatial distribution of individually sponds to the “Color Code 3” (Stiner et al., plotted larger bones with impact marks 1995, p. 226), which mean they are fully car- and small bone chips from GB show differ- bonized. Burnt bone concentrations within ent concentration (marked in red; Fig. 6) the three sublevels analysed here cluster dif- indicating varied density areas. In the lower ferently, while showing some degree of spa- sublevel, 16.2 bones with impact marks tial overlap. High concentrations of burned concentrate in subsquare J3-D, while more bones are found in the lower sublevel 16.2 in loosely dispersed throughout the remaining subsquares J4-D, J3-B,D; in the middle sub- excavated squares. In sublevel 16.1, bones level 16.1 in subsquare I4-B and subsquares with impact cluster within squares J3 and J4-B,C,D; and finally in the upper sublevel 15 H4 as well as subsquare H3-C. In sublevel in I4-B,C and J4-D (Fig. 5). 15, bones with impacts marks were found in

Figure 5: The small refuse burnt bone distribution varies between sublevels 15, 16.1 and 16.2. Amounts of burnt bone remains constant in the northeastern edge of the excavation areas (square J4), which may be associated with a hearth.

Figure 6: The two components, impact marks on bones and bone flakes or chips, are com- bined to achieve the distribution of bone working areas contributing to marrow extracting or grease rendering activities. Individual faunal remains with impact marks are plotted indi- vidually. Based on the counts of bone flakes the subsquares were colour coded according to low and high densities, from blue to red respectively. Geiling et al: A Spatial Distribution Study of Faunal Remains from Two Lower Art. 4, page 9 of 16 Magdalenian Occupation Levels in El Mirón Cave, Cantabria, Spain squares H3 and H4 as well as in the south- we will discuss our first insights into the eastern edge of the excavation area (square activity areas of two LCM layers in El Mirón J2). Bone flakes and chips from GB show a cave, based on spatial distribution analysis of different spatial pattering. In Sublevel 16.2, faunal remains. these find categories clustered within subs- This paper analysed general point pat- quares J3-D, J2-D and square I3. The super- terns using kernel density and created heat- imposed sublevel 16.1 shows high amounts maps to show clustering of individual finds. of bone chips in subsquares J3-B, H4-A, -C, The faunal distributions in levels 15 and 16 -D and H3-C, while within sublevel 15 chips inside the Cabin area are not homogeneous concentrate in subsquare J2-D. Within the (Fig. 4), according to ‘first-order’ characteris- sublevels analysed here the bone flakes con- tics (Bevan et al., 2013). The excavation area centrations and bones with impact-marks represents a spatial heterogeneity, because coincide in some areas. In sublevels 16.2 and the southern excavation border is the cave 16.1 these categories overlap in square J3; wall. Due to accumulation principles, items within sublevel 16.1, additional overlap was do cluster along vertical borders (Harris identified in squares H3 and H4. In sublevel 1989). This might be the reason for accumu- 15, no overlap was detected. lation clusters along the southern squares, but does not explain varied point intensities Discussion in other areas. Low point intensities next to The zooarchaeological and taphonomic stud- hot spots are visible along the western and ies of dense bone accumulations within El eastern excavation borders. They were pre- Mirón cave revealed excellent conservation sumably created due to ‘edge effects’ of point conditions. Particularly, good bone pres- distribution while using kernel density anal- ervation indicates little post-depositional ysis (Orton, 2004). In any event, these pat- disturbance. The existence of young animal terns do not affect the question addressed remains (very porous, spongy, tiny bones) here concerning general bone distribution demonstrate quite limited non-cultural for- by ungulate species. Our results explored mation processes. During excavations no whether the two most hunted animal species, significant sediment movement, pressure or red deer and ibex, overlapped in their spatial wash-outs were observed, phenomena which distribution patterns (Fig. 4). Their almost would have suggested a disturbance of origi- complete coincidence implies that both nal archaeological layers (Gonzalez-Morales species (with similar anatomies but moder- and Straus, 1996–2011; Straus and Gonzalez- ately different average sizes) were butchered, Morales, 2012b). These results demonstrate exploited and discarded in the same way and the existence of an intact archaeological therefore suggests a uniformity of accumu- stratigraphy and LCM strata in El Mirón that lation processes. We have shown that non- are very suitable for investigating the spatial cultural formation/destruction processes distribution pattern of finds, from which were very limited relative to cultural ones, evidence of activity areas can be inferred. leading to confidence in the following dis- Additionally, cultural formation processes cussion about activity areas within Levels 15 can be used to enhance interpretations of and 16 of the El Mirón Outer Vestibule exca- human behaviour. Due to the problems vation area, with the obvious caveat that this archeologists face when attempting analysis 9.25 m² area represents a very small portion of very thick cave deposits (the palimpsest of the whole ca. 300 m² habitable surface of problem), spatial distribution analysis can the vestibule. simplify the data, enabling one to recognize Activity zones can be distinguished within spatial patterns among archaeological find the sublevels of the LCM palimpsest horizon categories that are not visible with the naked in El Mirón cave (Fig. 7), in particular hearth, eye. Consequently, in the following sections working and dumping areas can be seen. Art. 4, page 10 of 16 Geiling et al: A Spatial Distribution Study of Faunal Remains from Two Lower Magdalenian Occupation Levels in El Mirón Cave, Cantabria, Spain

Figure 7: There exists a spatial sorting of the bone types: long bones fragments cluster in two areas as opposed to smaller bones (short or irregular) and teeth. The spatial interpretation of activity zones inside the LCM camp in El Mirón cave refers to bone working, dropping, walking, tossing and dumping zones. A hearth was presumably located around square J4. A bone processing area around square J3 appears to have played a central role in the organ- ization of reconstructed activities.

The spatial data of the faunal remains per- 2004). Whether the burnt bone fragments mits archaeologists to identify activity areas, at El Mirón Level 15 and 16 were simply dis- comparable to ethnographic examples of dis- posed of into the fires or served as a second- posal modes in bone assemblages (Bartram ary fuel source (wood was relatively scarce et al., 1991; Binford, 1978). during Oldest Dryas in this region) remains The distribution of small burnt bones open for discussion. concentrates around square J4. The spatial Intentional bone cracking activity areas patterning of burned remains may indicate have been identified in square J3, in both the location of a hearth in the northeast sublevels 16.2 and 16.1. Bones were delib- of the excavation area. In addition to the erately fragmented, probably in the course burnt material, ashy dark, organic rich sedi- of marrow extraction and preparation for ments were excavated in this square of the grease rendering activities (Janzen et al., site (Gonzalez-Morales, Straus and Student 2014). Three criteria are necessary for the Excavators unpub Field notes). The bones identification of bone grease rendering at were possibly used to fuel the fires at the prehistoric sites: stone anvils to break bones site. Spongy, cancellous bones are better prior to cooking, fire-cracked rocks indica- fuel, while dense long bone shaft fragments tive of stone boiling and highly fragmented are unsuitable as fuel, given that they do not spongy bones (Binford, 1978; Stiner, 2003). burn reliably (Bosch et al., 2012; Costamagno Stone-boiling has been found in et al., 2006; Morin, 2010; Théry-Parisot, the El Mirón LCM deposits, which indicates 2002; Théry-Parisot et al., 2005; Villa et al., cooking activities in the same area and time Geiling et al: A Spatial Distribution Study of Faunal Remains from Two Lower Art. 4, page 11 of 16 Magdalenian Occupation Levels in El Mirón Cave, Cantabria, Spain period (Nakazawa et al., 2009). Despite the rearticulated within the walking zone. fact that the quartzic cobbles were Additionally, narrowing down excavation probably used as percussion stone or anvils spits to check for bone refits might help the before being heated, no other specific stone identification of activity areas within such anvils for the purpose of bone cracking were thick archaeological cave deposits. We are identified (Straus and Gonzalez-Morales, aware that present bone refits are the only 2007b; Straus and González-Morales, 2012a). way with which to prove the existence of liv- A drop zone (around 1m in radius), located ing floors in LCM layers of El Mirón, since the on the peripheries of the working zone in associated lithic artefact assemblages have square J3 was determined, based on larger not yet been studied from this perspective. bone leftovers with impact marks. The work- The occupations in Level 15 and 16 in El ing area was possibly a central locus of bone Mirón cave represent the uppermost layers processing activity close to the fireplace and and therefore the end of the LCM occupa- the dropping zone. tion in the excavated Cabin area. The iden- Along the surroundings of the bone- tification of specific activities and aspects processing zone, within a two meter of campsite organisation reveals insights radius, the high density of long bone items into the behaviour of LCM hunter-gatherers. decreases and mostly teeth and other bone The LCM cultural period initiated quickly types such as short or irregular bones are changing , settlement patterns, represented. The area may be interpreted subsistence strategies, and artistic activities as a walking corridor as it contains fewer (Straus 1992). Whether the El Mirón accumu- long bone fragments­ (squares I2, I3 and I4). lations are the results of one long or (more Further away (at more than a 2 m radius) likely) several relatively short, but repeated, from the working zone, a dense accumula- occupations remains an open question but tion of faunal material with anthropogenic they represent a change in resource usage modifications is interpreted as a toss zone in the cave relative to the levels (squares H2, H3 and H4). therein. While red deer and ibex meat were In subsquare H3-A across sublevel 16.1 definitely common components of the LCM another small cluster of small bone refuse diet (see also Marín-Arroyo and Geiling, fragments, burnt bones and bones with 2015), additional regular extractions of bone impact marks was found. Given the over- marrow and grease imply a more intensive all thickness (1–2 cm; one spit) and dense exploitation of these animals. The recon- accumulation, this area could be inter- structed activities also indicate a continu- preted as a dumping zone (O’Connell et al., ous/repeated use of these working areas, 1991; Schiffer, 1996). Ethnographical com- while spaces were maintained/reused. parisons have shown that hearths and adja- Examples include the bone working zone cent areas were cleaned only when primary overlapping with cobbles presumably used refuse densities reached unpleasant levels. several times to open bone marrow cavities, After removing the larger pieces, only the and the evidence for dumping events indicat- smaller fragments remain in place (Schiffer, ing cleaning procedures on site and hence 1996). Cleaning processes might be neces- some degree of longevity and/or repetition sary during long or frequent occupations, as of use of this space. The thickest and densest is implied for the LCM El Mirón residential LCM layer is the underlying Level 17, which hub campsite. is currently being studied by the first author The distribution of activities zones within () and by L.M. Fontes (lithic artefacts). the LCM layers of El Mirón is supported by The abrupt end of these occupations after refittings of the faunal material. The num- Level 15 might reflect the abandonment of ber of bone refits increase within the toss the cave by LCM hunter-gatherer groups. In or drop zones, while teeth are commonly contrast, despite higher sedimentation rates, Art. 4, page 12 of 16 Geiling et al: A Spatial Distribution Study of Faunal Remains from Two Lower Magdalenian Occupation Levels in El Mirón Cave, Cantabria, Spain the overlying stratigraphic layer (Level 14) in currently housed at the Instituto Internacional the El Mirón Outer Vestibule representing de Investigaciones Prehistóricas de Cantabria Middle Magdalenian occupations, displays very (IIIPC) and the Museo de Prehistoria low archaeological and faunal densities com- y Arqueología (MUPAC) Santander. J.M. Geiling pared to those described here (Marín-Arroyo, is grateful to L.G. Straus and A.B. Marín- 2010; Straus and González Morales 2012a). Arroyo for useful comments on earlier drafts of this manuscript, to Y.H.Hilbert Conclusions for insightful discussions on the subject The use of GIS analysis on faunal remains in and to J. Jones and L.G.Straus for revising Palaeolithic occupation horizons can yield the English grammar spelling. This study positive results when dealing with large was partially presented by J.M.Geiling at amounts of quantifiable data. The creation the Postgraduate Zooarchaeology Forum of plots is the first step, while additional sta- (PZAF) 2014 in London. J.M.Geiling is tistical analysis can be added. Density cluster funded by a predoctoral grant (Formación and heatmaps not only enhance archaeologi- de Personal Investigador: BES-2013-063309) cal interpretation, but also produce results from the Spanish Ministry of Economy and that can be compared across archaeological Competitiveness. and ethnographical sites. This useful tool is of particular interest in analysing dense Competing Interests occupational layers such as the LCM depos- The authors declare that they have no com- its in the Cantabrian region. In sum, through peting interests. the additional inclusion of GB data, it was possible to reconstruct activities in the LCM References levels 15 and 16 in El Mirón cave. This case Alperson-Afil, N, Sharon, G, Kislev, M, study has shown that a range of spatial data Melamed, Y, Zohar, I, Ashkenazi, S, can be synthesized by combining different Rabinovich, R, Biton, R, Werker, E, faunal categories from Individual and GB Hartman, G, Feibel, C and Goren-Inbar, N data. These interpretations of intense LCM 2009 Spatial Organization of Homi- occupations are made possible by spatial dis- nin Activities at Gesher Benot Ya’aqov, tribution plots showing different anthropo- . Science, 326(5960): 1677–1680. genic activities. DOI: http://dx.doi.org/10.1126/science. Activity areas identified at El Mirón are a 1180695 grease-rendering working zone for bone pro- Audouze, F and Enloe, J G 1997 High cessing, a drop zone, a walking zone, hearths, ­resolution archaeology at Verberie: limits a dumping area and a toss zone. The bone- and interpretations. World Archaeology, processing zone had a central role in the 29: 195–207. spatial organization of this area of the site. Bailey, G 2007 Time perspectives, palimp- This specific activity produces a large quan- sests and the archaeology of time. Jour- tity of bone debris, which in turn is visible nal of Anthropological Archaeology, 26(2): in the high-density LCM layers found across 198–223. Northern Spain and makes it archaeologi- Bartram, L E, Kroll, E M and Bunn, H T cally tangible. Whether these thick deposits 1991 Variability in camp structure and represent specific cultural activities of the bone food refuse patterning at Kua San LCM remains to be seen and will be the sub- hunter-gatherer camps. In: The interpre- ject of future research. tation of archaeological spatial pattern- ing. Springer, pp. 77–148. Acknowledgements Baxter, M J, Beardah, C C and Wright, R V S Faunal material from El Mirón was made 1997 Some Archaeological Applications available for study by J.M. Geiling by L.G. of Kernel Density Estimates. Journal of Straus and M.R. González-Morales and is Archaeological Science, 24(4): 347–354. Geiling et al: A Spatial Distribution Study of Faunal Remains from Two Lower Art. 4, page 13 of 16 Magdalenian Occupation Levels in El Mirón Cave, Cantabria, Spain

Bevan, A, Crema, E, Li, X and Palmisano, A Fladerer, F A, Salcher-Jedrasiak, T A and 2013 Intensities, interactions and Händel, M 2002 Hearth-side bone ­uncertainties: some new approaches to assemblages within the 27 ka BP Krems- archaeological distributions. In Compu- Wachtberg settlement: Fired ribs and tational Approaches to Archaeological the mammoth bone-grease hypothesis. Spaces. Walnut Creek: Left Coast Press, Quaternary International, 115–133. DOI: pp. 27–52. http://dx.doi.org/10.1016/j.quaint. Binford, L R 1978 Dimensional Analysis of 2012.06.030 Behavior and Site Structure: Learning Geiling, J M 2014 Taphonomic analysis of from an Eskimo Hunting Stand. American Lower Cantabrian Magdalenian deposits Antiquity, 43(3): 330–361. DOI: http:// from El Mirón cave (Cantabria, Northern dx.doi.org/10.2307/279390 Spain). ORAL COMMUNICATION. In: Post- Blankholm, H P 1991 Intrasite spatial graduate Zooarchaeology Forum (PZAF). analysis in theory and practice. Aarhus London. Universitetsforlag. Geiling, J M and Marín-Arroyo, A B Bosch, M D, Nigst, P R, Fladerer, F A and 2015 Spatial distribution analysis of Antl-Weiser, W 2012 Humans, bones the Lower Magdalenian human burial and fire: Zooarchaeological, tapho- in El Mirón Cave (Cantabria, Spain). nomic, and spatial analyses of a Gravet- Journal of Archaeological Science, 60: tian mammoth bone accumulation at 47–56. DOI: http://dx.doi.org/10.1016/j. Grub-Kranawetberg (Austria). Quaternary jas.2015.03.005 International, 252: 109–121. DOI: http:// Gilead, I 2002 Too many notes? virtual dx.doi.org/10.1016/j.quaint.2011.08.019 recording of artifacts provenance. In Clarke, D L 1977 Spatial archaeology. Lon- Virtual­ archaeology: proceedings of the don: Academic Press. VAST Euroconference, Arezzo 24–25 Cordain, L, Miller, J B, Eaton, S B, November 2000, pp. 41–44. Mann, N, Holt, S H and Speth, J D González Echegaray, J and Freeman, L G 2000 Plant-animal subsistence ratios 1992 Las Excavaciones de la Cueva del and macronutrient energy estimations Juyo (Cantabria). Kobie. in worldwide hunter-gatherer diets. The Gonzalez-Morales, M R and Straus, L G American journal of clinical nutrition, 1996–2011 Field notes. Unpublished 71(3): 682–692. documents, University of Cantabria. Costamagno, S 2013 Bone Grease ­Rendering Gonzalez-Morales, M R and Straus, L G in Contexts: The Case of 2014. Cueva del Mirón (Ramales de la Noisetier Cave (Fréchet-Aure, Hautes- Victoria, Cantabria). In and Pyréné es, ). In: Zooarchaeology and hunter-gatherers in Iberia and the Modern Human Origins, Vertebrate Paleo- strait: the current archaeological biology and Paleoanthropology. Springer record. Burgos: Universidad de Burgos. Netherlands, Dordrecht, pp. 209–225. González-Morales, M R, Straus, L G and Costamagno, S, Théry-Parisot, I, Brugal, J-P Marín-Arroyo, A B 2006 Los omópla- and Guibert, R 2006 Taphonomic tos decorados magdalenienses de la ­consequences of the use of bones as fuel. Cueva del Mirón (Ramales de la Victoria, Experimental data and archaeological Cantabria) y su relación con las Cueva del applications. In Biosphere to Lithosphere Castillo, Altamira y El Juyo. Zona arque- New studies in vertebrate taphonomy. ológica, (7): 483–494. Oxbow Books, pp. 51–62. Gonzalez-Morales, M R, Straus, L G and Ferring, C R 1984 Intrasite spatial pattern- Student Excavators unpub Field notes – ing: Its role in settlement-subsistence sys- Primary Excavation documentation: tems analysis. In: Intrasite spatial analysis “Meter-square spit forms” from El Mirón in archaeology, pp. 116–126. Cave (1996–2013). Art. 4, page 14 of 16 Geiling et al: A Spatial Distribution Study of Faunal Remains from Two Lower Magdalenian Occupation Levels in El Mirón Cave, Cantabria, Spain

Goren-Inbar, N and Alperson-Afil, N 2010 macromammal remains stratigraphically The Acheulian site of Gesher Benot Ya’aqov associated with the Magdalenian human Ancient flames and controlled use of fire, burial in El Mirón Cave (Cantabria, Spain). Dordrecht, London: Springer. Journal of Archaeological Science, 60: Harris, E C 1989 Principles of archaeologi- 75–83. DOI: http://dx.doi.org/10.1016/j. cal stratigraphy, 2nd edition. Academic jas.2015.03.009 Press, London. McKellar, J A 1983 Correlations and the Henry, D 2012 The palimpsest problem, Explanation of Distributions. Arizona hearth pattern analysis, and Middle Pale- Anthropologist, 4(0): 2–5. olithic site structure. Quaternary Interna- Morin, E 2010 Taphonomic implicatons of tional, 247: 246–266. the use of bone as fuel. In: The taphon- Hodder, I and Orton, C 1976 Spatial analysis omy of burned organic residues and com- in archaeology. New Studies in Archaeol- bustion features in archaeological contexts ogy. Cambridge. (Proceedings of the Round Table, Valbonne, Janzen, A, Reid, R E B, Vasquez, A and May 27–29 2008, CEPAM). Gifford-Gonzalez, D 2014 Smaller frag- Nakazawa, Y, Straus, L G, González- ment size facilitates energy-efficient bone Morales, M R, Solana, D C and Saiz, J C grease production. Journal of Archaeo- 2009 On stone-boiling technology in the logical Science, 49: 518–523. DOI: http:// Upper : behavioral implications dx.doi.org/10.1016/j.jas.2014.06.004 from an Early Magdalenian hearth in El Leroi-Gourhan, A and Brézillon, M 1966 Mirón Cave, Cantabria, Spain. Journal of L’habitation magdalénienne n° 1 de Pince- Archaeological Science, 36(3): 684–693. DOI: vent près Monterau (Seine-et-Marne). http://dx.doi.org/10.1016/j.jas.2008.10.015 Gallia préhistoire9, pp. 263–385. DOI: O’Connell, J F, Hawkes, K and Jones, N B http://dx.doi.org/10.3406/galip.1966.1264 1991 Distribution of refuse-producing Lyman, R L 1994 Vertebrate taphonomy, activities at Hadza residential base camps. Cambridge University Press. The interpretation of archaeological Malinsky-Buller, A, Hovers, E and Marder, O spatial patterning. Springer, pp. 61–76. 2011 Making time: “Living floors”, “pal- Oron, M and Goren-Inbar, N 2014 Mousterian impsests” and site formation processes – intra-site spatial patterning at Quneitra, A perspective from the open-air Lower Golan Heights. Quaternary International, Paleolithic site of Revadim Quarry, Israel. 331: 186–202. Journal of Anthropological Archaeology, Orton, C 2004 Point pattern analysis 30(2): 89–101. DOI: http://dx.doi.org/ revisited. Archeologia e Calcolatori, 15: 10.1016/j.jaa.2010.11.002 299–315. Manne, T, Stiner, M C and Bicho, N F Outram, A K 1998 The identification and 2006 Evidence for bone grease render- Palaeoeconomic context of prehistoric ing during the at Vale bone marrow and grease exploitation. Boi (Algarve, ). Animais na Pré- Department of Archaeology, University of História e Arqueologia da Península Ibé- Durham. rica. Actas do IV Congresso de Arqueología Reitz, E J and Wing, E R 2008 Zooarchae- Peninsular, 3: 145–158. ology, Cambridge; New York: Cambridge Marín-Arroyo, A B 2010 Arqueozoología en University Press. el Cantábrico Oriental durante la tran- Schiffer, M B 1983 Forward to McKellar Paper. sición Pleistoceno/Holoceno. La cueva del Available at: http://arizona.openrepository. Mirón, Serie Tesis Doctorales, 2. Santander: com/arizona/handle/10150/112175. PUbliCan. Schiffer, M B 1996 Formation Processes of Marín-Arroyo, A B and Geiling, J M the Archaeological Record 1st Edition. Salt 2015 Archeozoological study of the Lake City: University of Utah Press. Geiling et al: A Spatial Distribution Study of Faunal Remains from Two Lower Art. 4, page 15 of 16 Magdalenian Occupation Levels in El Mirón Cave, Cantabria, Spain

Silverman, B W 1986 Density Estimation the site and its Holocene archaeological for Statistics and Data Analysis, Springer- record, Albuquerque: University of New Science+Buisness Media. B.V. Mexico Press. Stiner, M C 2003 Zooarchaeological Straus, L G and González-Morales, M R ­Evidence for Resource intensification 2012b The Magdalenian settlement of Algarve Portugal. Promontoria, 1(1). the Cantabrian region (Northern Spain): Stiner, M C, Kuhn, S L, Weiner, S and The view from El Miron Cave. Quaternary Bar-Yosef, O 1995 Differential burning, International, 272–273: 111–124. recrystallization, and fragmentation of Straus, L G and Gonzalez-Morales, M R archaeological bone. Journal of Archaeo- 2015 The Magdalenian Human Burial logical Science, 22(2): 223–237. of El Mirón Cave (Ramales de la Vic- Straus, L G 1979 : a palaeoanthropo- toria, Cantabria, Spain): Introduction, logical resource. World Archaeology, 10: Background, Discovery and Context. 331–339. Journal of Archaeological Science, 60: Straus, L G 1982 Carnivores and cave sites 1–9. DOI: http://dx.doi.org/10.1016/j. in Cantabrian Spain. Journal of Anthropo- jas.2015.02.018 logical Research, 38(1): 75–96. Straus, L G, Gonzalez-Morales, M R and Straus, L G 1992 Iberia before the Iberians: Carretero, J-M 2011 Lower Magdale- the of Cantabrian nian secondary human burial in El Mirón Spain, University of New Mexico Press. Cave, Cantabria, Spain. Antiquity, (85): Straus, L G 2008 Palimpsest, assemblages, 1151–1164. phases & facies: the conundrum of Théry-Parisot, I 2002 Fuel Management ­taxonomic generalization from particular (Bone and Wood) During the Lower Aurig- archaeological cases. In PROCEEDINGS nacian in the Pataud (Lower OF THE XV WORLD CONGRESS (LIS- Palaeolithic, Les Eyzies de Tayac, Dor- BON, 4–9 SEPTEMBER 2006). Oxford: dogne, France). Contribution of Experi- Archaeopress. mentation. Journal of Archaeological Straus, L G and González-Morales, M R Science, 29(12): 1415–1421. DOI: http:// 2003 El Miron Cave and the 14C Chro- dx.doi.org/10.1006/jasc.2001.0781 nology of Cantabrian Spain. Radiocarbon, Théry-Parisot, I, Costamagno, S, Brugal, J-P, 45(1): 41–58. Fosse, P and Guilbert, R 2005 The use Straus, L G and Gonzalez-Morales, M R of bone as fuel during the palaeolithic, 2007a Further Radiocarbon Dates for experimental study of bone combustible the Upper Paleolithic of El Mirón Cave properties. In: The zooarchaeology of fats, (Ramales de la Victoria, Cantabria, Spain). oils, milk and dairying. Oxford: Oxbow Radiocarbon, 49(3): 1205–1214. Books. Straus, L G and González-Morales, M R Villa, P, Castel, J-C, Beauval, C, Bourdillat, V 2007b Early Tardiglacial human uses of El and Goldberg, P 2004 Human and car- Mirón Cave (Cantabria, Spain). BAR Inter- nivore sites in the European Middle and national Series, 1655: 83–94. Upper Paleolithic: Similarities and dif- Straus, L G and Gonzalez-Morales, M R ferences in bone modification and frag- 2010 The Radiocarbon Chronology of El mentation. Revue de Paléobiologie, 23(2): Mirón Cave (Cantabria, Spain): New Dates 705–730. for the Initial Magdalenian Occupations. Whallon, R 1984 Unconstrained clustering Radiocarbon, 52(1): 33–39. for the analysis of spatial distributions in Straus, L G and Gonzalez-Morales, M R archaeology. In: Intrasite spatial analysis 2012a El Miron Cave, Cantabrian Spain: in archaeology. Cambridge. Art. 4, page 16 of 16 Geiling et al: A Spatial Distribution Study of Faunal Remains from Two Lower Magdalenian Occupation Levels in El Mirón Cave, Cantabria, Spain

How to cite this article: Geiling, J M, Straus, L G, González-Morales, M R and Marín-Arroyo, A B 2016 A Spatial Distribution Study of Faunal Remains from Two Lower Magdalenian Occupation Levels in El Mirón Cave, Cantabria, Spain. Papers from the Institute of Archaeology, 26(1): Art. 4, pp. 1–16, DOI: http://dx.doi.org/10.5334/pia-477

Submitted: 08 December 2014 Accepted: 02 November 2015 Published: 31 August 2016

Copyright: © 2016 The Author(s). This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International License (CC-BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. See http://creativecommons.org/licenses/by/4.0/.

Papers from the Institute of Archaeology is a peer-reviewed open

access journal published by Ubiquity Press. OPEN ACCESS